How to Forecast Community Annoyance in
Planning Noisy Facilities
R. Guski
Ruhr University Bochum, Germany
When planning the development or reduction of large traffic facilities, acoustic calculation
procedures are used to forecast the noise load in the affected residential areas. Then, existing
dose/response relationships for steady state situations are used to predict noise effects in future
years. Planners often assume that (1) noise annoyance reactions of residents do not change
over the years, and (2) annoyance is not affected by the change itself. Both of these assumptions
are questioned in this paper, and a procedure for estimating future annoyance in changed
noise situations is proposed. This includes the analysis of possible statistical trends of the
annoyance reactions over the years - even for steady-state noise loads, and with changing state
situations, the effects of the change should also be accounted for.
Keywords: Community noise; Annoyance; Dose/Response relationships; Trends over time; Effects of
Noise Change.
Introduction
When planning the development of traffic
facilities like airports, roads, and railway tracks,
acoustic calculation procedures are used in order
to forecast the noise load in the vicinity of the
new or changed facility, especially in residential
areas nearby. With large developments, the delay
between the planning and the opening of the new
or changed facility may take several years. In
forecasting the effects of noise in the future, it is
necessary to know (a) whether dose/response
relationships established some years ago are
valid today, and will be valid in the future, and
(b) whether changing a noise situation (e.g.
opening a new airport runway) has special
annoying effects on residents which are not seen
in steady state noise situations. Planners often
assume that (1) noise annoyance reactions of
residents do not change over the years, and (2)
annoyance is not affected by the change itself.
Both of these assumptions can be questioned.
Does annoyance change over the years?
In many countries, large-scale traffic or
industrial
developments
require
an
Environmental Impact Statement (EIS) before
constructions can start. The EIS requires detailed
analyses of any action that may significantly
affect the quality of the environment, and the
complete statement comprises many volumes.
Two parts of the EIS should contain predictions
of the noise situation in residential areas after
opening the new (or changed) development: one
part should forecast the acoustic situation (e.g.,
noise contours in the residential neighborhood),
and the other part should forecast the health and
noise annoyance situation. Both of these tasks
require considerable skills and assumptions; for
instance, predicting the acoustic situation after
opening a new airport runway in 10 years
requires good knowledge (or, at least a good
guess) of the mix of aircraft types at this airport
in 10 years, their relative contribution to the total
noise load, their operating characteristics, flight
paths, and so on. Forecasting the health and
noise annoyance situation after opening the new
runway is not less complicated, because data on
trends of noise annoyance and trends of noise
related health over time is rare, and data on the
effects of changing the noise levels is partially
conflicting.
When planning for new noisy developments or
noise abatement programs, planners and
consultants like to use published dose-effect
Noise & Health 2004, 6;22, 59-64
Figure 1. Dose/response relationships for three Swiss airports (1971/1991). Please note that the
disturbance questions are not identical.
relationships like the well-known curves from
the TNO data set (e.g. Miedema and Vos 1998).
This data set comprises annoyance and
disturbance data from 55 international
systematic field studies which have taken place
between the late 50s and the early 90s of the last
century. This data set is very comprehensive, but
we should not forget that the mean age of data is
23 years. In other words: the average vintage of
data is 1980. In the mean time, the structure of
the noise load has changed – even with equal
energy noise levels – for some noise sources,
especially for aircraft and road traffic noise: the
average noise level of individual vehicles
decreased, but the number of events increased,
and the duration of quiet periods decreased.
When planners use dose/response data from
1980 in order forecast noise effects in 2010, they
should make certain that the dose/response
relationship between noise level and annoyance
did not change in the mean time.
keep the noise sources (and their respective
composition) constant over several years.
Unfortunately, no such study does exist, and it is
questionable whether such a study could be
performed at all, because residential areas
without any change in noise levels are rare. A
less ideal study design would include repeated
measurements over several years in several
residential areas which underwent not more than
the typical gradual change in noise levels which
can be observed all over the industrialized world.
In this case, both noise exposure and reaction
variables should be measured repeatedly in the
same way, such that repeated dose/response
relationships could be established. The updated
catalogue of social surveys on noise (Fields
2001) mentions one French study and a Swedish
one (Jonsson, Sörensen, Arvidsson, and
Berglund 1975). The French reports are not
available any more, and the Swedish report is
rather short and does not give details of the
measurement procedures and of dose-response
In order to test for long-term developments of relationships; however, it does say that during
noise annoyance in residential areas, an ideal the eight years between 1963 and 1971, the
study design would include repeated average fraction of “rather or very disturbed”
measurements (using exactly the same questions persons increased about 3 per cent.
and measurement procedures) of reaction
variables over several years in several areas The Fields catalogue does not mention the partial
comprising a range of noise levels. The areas repetition of the initial Swiss study on aircraft
should not change over time, i.e. they should noise (Grandjean et al. 1973, containing data
60
Figure 2. Percentage of Highly Annoyed residents at Düsseldorf Airport 1987-1993. After Kastka et al.
(1995).
from 1971) which has been reported by Oliva
(1998). This is by no means an ideal comparison
study: The three airports (Basel, Geneva and
Zurich) underwent a considerable change, the
aircraft noise calculation procedure changed
from NNI to Leq, and the questions to
respondents changed too. The old study asked
for “disturbances due to aircraft noise” without
mentioning a specific reference to the location
(e.g., inside/outside the house). The later study
(containing data from 1991) posed different
questions for different locations, and Oliva
(1998) thinks that responses to “outside”
questions are comparable with responses to the
old questions without specific location reference.
If we agree on this assumption, there is little
difference between dose/response relationships
for aircraft noise in Switzerland: Only above
NNI > 30, there is a small increase in the
percentage of disturbed people within 20 years
(see Figure 1). This small (and statistically
insignificant) increase gets some added value if
we consider that sound insulation had been
installed between the two studies in residential
areas with noise loads greater than NNI=42.
Sound abatement programs are usually expected
to decrease disturbance and annoyance
judgments, and if the percentage of highly
disturbed persons was the same or slightly higher
in 1991 as compared to 1971, this could mean (a)
that the sound abatement program was not
effective, or (b) that aircraft noise annoyance
increased in the mean time.
There is another example from aircraft noise:
Kastka et al. (1995) report repeated surveys in
residential areas at Düsseldorf Airport
(Germany) in 1987 and 1993. This comparison is
of particular interest, because LAeq for daytime
landings decreased about 2.1 dB(A) between
19987 and 1993, but annoyance judgments
increased (see Figure 2). Global annoyance was
measured on a 7-point scale, and those persons
who chose one of the upper three points on this
scale were counted as “Highly Annoyed”. The
fraction of highly annoyed residents increased
during the 6 years about 20 percent.
Unfortunately, the data do not allow for
calculating a statistical trend over time.
The least ideal approach for answering the
question of an annoyance trend over time is to
use data sets from different studies performed at
different times in different countries. This
approach ignores systematic differences between
studies, but since most comparisons of
dose/response relationships do so, we could as
well take the dose/response relationships for
aircraft noise and road traffic noise given by
Miedema and Vos (1998), select data for a
61
Figure 3. Noise levels for a constant proportion (25%) of highly annoyed residents. Data from Miedema
and Vos (1998).
constant fraction of highly annoyed residents,
and rearrange the results according to age of
publication. In doing so, we find a decrease of
the day/night level necessary for evoking a
constant percentage of 25% respondents being
highly annoyed by aircraft noise, and
inconsistent results with respect to road traffic
noise (Figure3).
between 1970 and 1983, and kept a rather
constant level afterwards. Although there is
considerable variation between different studies
within the same year, the nonlinear regressions
fit significantly to the data points (r2 = 0.38 for
aircraft, and r2 = 0.36 for road traffic): The
annoyance change for aircraft noise is equivalent
to 6 dB DNL between 1965 and 1985. When
planning a noise load that will be effective in
It should be noted that the regression lines in
several years, it may be necessary to calculate
Figure 3 are calculated without weighting the
the annoyance trend over the last 20 years, and
number of respondents in each of the studies
provide for a potential change of annoyance in
included. This may bias the result, and a closer
the future – e.g., by extrapolating the statistical
look into the data may be necessary, but the first
trend. In the case of aircraft noise, this would
impression is that the annoyance of residents
mean to adapt land use planning and noise
exposed to aircraft noise increased over the
abatement programs to the statistical trend of
years, while the road traffic annoyance increased
Figure 4. Schematic view of community annoyance changes with step changes of noise level.
62
community noise annoyance. (It should be noted
that no trend was estimated for railroad
annoyance, because the number of available data
is too small for this purpose).
(1997) assume that the “evolutionary” expansion
of Heathrow Airport (which mainly took place
between 1961 and 1965) contributed to the fact
that dose/response curves established 1967 were
almost the same as in 1961. On the other hand,
Annoyance in changing noise situations: the step change at Vancouver Airport 1996
When planning for a new or significantly provoked 41 % more highly annoyed residents in
changed noise situation in the future (e.g. one area which underwent an increase of 7 dB
opening or closing a road, or opening or closing (DNL) from one day to the next (Fidell et al.
an airport runway), an additional effect should be 2002, see Figure 5).
taken care of: Residents react to the change of
the noise situation. When the noise situation is The amount of “overshoot” also depends on the
abruptly and permanently changed the amount of change at each location: Fidell et al.
annoyance of residents usually changes in a way 2002 report a follow-up study two years after the
that cannot be predicted by steady-state opening of a new runway at Vancouver Airport.
dose/response relationships (cf. Fidell et al. It turned out that in 3 residential areas which did
2002; Raw and Griffiths 1990): Most studies not undergo a measurable change in noise levels,
show an „over reaction“ of the residents, i.e., the percentage of highly annoyed residents
with an increase of noise levels, people are much stayed about the same as before (or even
more annoyed than would be predicted by decreased about 5%). In one area which
steady-state curves, and with a decrease of noise underwent a decrease of 1 dB, the percentage of
levels, people are much less annoyed (Figure 4). highly annoyed residents stayed about the same
It should be noted that the annoyance level as before, but in all areas which underwent an
changes already before the change of levels: increase of noise levels, the increase of the
Residents expecting an increase of levels react percentage of highly annoyed residents
more annoyed, and residents expecting a depended on the amount of increase in noise
decrease in levels react less annoyed than would levels (see Figure 5). It is uncertain whether
be predicted in the steady state condition future studies will support this clear change
(Hatfield et al. 1998, 2002).
effect, but forecasting annoyance in changed
noise situations should take care of overshoot
The amount of “overshoot” depends on the reactions.
abruptness of change: Horonjeff and Robert
Figure 5. Differences in the percentage of highly annoyed residents in relation to the difference of DNL
levels after a step change at Vancouver Airport in 1996 (after Fidell et al. 2002).
63
A last point should not be neglected: Almost all
papers on change effects assume that overshoot
reactions will decay with time (e.g., Fields et al.
2000). But we do not know how long this decay
will take, and which variables will contribute to
a rapid decay. Raw and Griffiths (1990) claim
that some effects can be seen up to 9 years after
the change. We can hypothesize that “soft noise
abatement
programs”
(like
planning
participation by residents, see Flindell and Witter
1999; Stallen 2000) will reduce the overshoot
reaction, because such an effect can be predicted
by a psychological stress theory. But long-term
data on changing noise situations is very rare.
Grandjean, E., Graf, P., Lauber, A., Meier, H.P. and Müller,
R. (1973). A survey of aircraft noise in Switzerland.
Proceedings of the International Congress on Noise a
Public Health Problem (Dubrovnik), pp. 645-659.
Conclusion
Predicting future annoyance should include
possible statistical trends of the annoyance
reactions over the years – even for steady-state
noise loads, and with changing state situations,
the effects of the change should also be
accounted for. This could mean that many
current impact assessments of future noise
situations underestimate the actual annoyance of
the residents – at least with respect to aircraft
noise.
Jonsson, E., Sörensen, S., Arvidsson, O. and Berglund, K.
(1975). Reliability of forecasts of annoyance reactions. A
study on exposure to noise and air pollution. Archives of
Environmental Health, 30, 104-106.
Correspondence Address
R. Guski
Fakult. f. Psychologie
Ruhr-Uni Bochum
D-44780 Bochum, Germany
Tel +49 234 322 2670, Fax +49 234 321 4308
References:
Fidell, S., Silvati, L. and Haboly, E. (2002). Social survey
of community response to a step change in aircraft noise
exposure. Journal of the Acoustical Society of America,
111, 200-209.
Fields, J.M., Ehrlich, G.E. and Zador, P. (2000). Theory
and design tools for studies of reactions to abrupt changes
in noise exposure. NASA / CR-2000-210280, Langley
Research Center, Hampton, Va.: NASA.
Fields, J.M. (2001). An updated catalog of 521 social
surveys of residents’ reactions to environmental noise
(1943-2000). NASA/CR-2001-211257, Washington, D.C.:
National Aeronautics and Space Administration.
Flindell, I.H. and Witter, I.J. (1999). Non-acoustical
factors in noise management at Heathrow Airport. Noise
and Health, 3, 27-44.
64
Hatfield, J., Job, R.F.S., Carter, N.L., Peploe, R., Taylor, R.
and Morell, S. (1998). Attitude-mediated reaction to noise
influences physiological responses: Evidence supporting
causality. Inter-Noise’98, Paper No. 440.
Hatfield, J., Job, S., Faunce, G., Carter, N., Peploe, P.,
Taylor, R. and Morrell, S. (2002). The effect of changed
levels at Sydney Airport on health outcomes II: The role of
anticipation and reaction. Forum Acusticum, Paper NOI05-006.
Horonjeff, R.D. and Robert, W.E. (1997). Attitudinal
responses to changes in noise exposure in residential
communities. Hampton, VA: NASA CR-97-205813.
Kastka, J., Borsch-Galetke, E., Guski, R., Krauth, J.,
Paulsen, R., Schuemer, R. and Oliva, C. (1995).
Longitudinal study on aircraft noise. Effects at Düsseldorf
Airport 1981 - 1993. Proceedings of the 15th International
Congress on Acoustics (ICA95), Trondheim, 4, 447-451.
Miedema, H.M.E. and Vos, H. (1998). Exposure-response
relationships for transportation noise. Journal of the
Acoustical Society of America, 104, 3432-3445.
Oliva, C. (1998). Belastung der Bevölkerung durch Flugund Straßenlärm. Eine Lärmstudie am Beispiel der
Flughäfen Genf und Zürich. Berlin: Duncker and
Humblot.
Raw, G.J. and Griffiths, I.D. (1990). Subjective response
to changes in road traffic noise: a model. Journal of Sound
and Vibration, 141, 43-54.
Stallen, P.J. (2000). New ways (and old theory) for abating
airport noise. InterNoise 2000, Paper IN2000/31.